Chapter 2

Oscillations in one-particle systems

 Chapter 2

Introduction The problem of a mass attached to a spring was discussed in

previous physics courses. As you remember, the mass
oscillates around the point where the spring is not stretched.
The reason we are reviewing this problem here is that we
want to study waves, which are the oscillations of a system of
many masses connected by many springs. As you will see,
the generalization to many masses is surprisingly straightfor-
ward once you understand the problem of a single oscillating
mass.
00000
11111
HookeÕs law
11111
00000
Consider the mass attached to the spring in Fig. 1

11111
00000
11111
00000
11111
00000
11111
00000 K m
00000
11111
11111
00000
11111
00000
00000
11111
11111
00000
x

00000
11111 0
Figure 1 A mass connected to a spring is the typical example of
HookeÕs law.

It is known that the force exerted by the spring on the mass

is proportional to the displacement of the mass, namely F =
-Kx. The constant K, known as the spring constant, measures
how strong that force is. The minus sign in the expression
for the force is of fundamental importance. It means that
the force opposes the displacement, ultimately leading to
oscillations. Without the minus sign in the force, an object
displaced by any amount would never return to the starting
point.

The above force is an example of HookeÕs law. You may

wonder what is the origin of HookeÕs law. In the previous
chapter, we stated that there are only a few fundamental
forces in Nature. We did not mention HookeÕs law. This

phy 241 MenŽndez 24

 Chapter 2

implies that this law must be a special case of one of the

fundamental interactions. Let us consider what happens
when we stretch a spring, or, for that matter, any object.
Clearly, the separation between atoms, and, consequently,
their mutual forces, will change. The forces between atoms
are electrical in nature. The gravitational interaction between
the atoms of any human-size object is negligible, and the
strong nuclear force is not involved, because the distance
between protons and neutrons inside the nuclei does not
change. It is only the separation between atoms which
increases, and this must affect the Coulomb interaction
between electrons and nuclei. But the Coulomb interaction
is inversely proportional to the square of the separation
between charges. This makes it hard to understand HookeÕs
law, which states that the force is proportional to the separa-
tion. The key to understanding HookeÕs law is to realize that
a given atom interacts with all its neighbors. When a solid
object is formed, the atoms will look for the positions that
minimize their energy. In other words, the atoms will sit at a
position where their potential energy has a minimum. Near
a minimum, the potential energy can always be written as E p
= E 0 + cx2 + dx3 + fx 4 + ... There can be no linear terms of
the form bx, because the force must be zero at the equilibri-
dE p
um position. (Remember that F = − ). For x small, the
dx
cubic and higher terms will be negligible, so that we end up
with a potential energy that has a parabolic form. However,
you remember that this is exactly the case for a spring that
obeys HookeÕs law: its potential energy is given by E p =
1/ Kx2. Therefore, it is not surprising that most objects obey
2
HookeÕs law. However, there is a limitation: The displace-
ment from the equilibrium position must be small, other-
wise, the cubic and higher terms in the expansion of the
potential energy become important. This can be observed
experimentally: if you apply a very strong stretching force to
an object, eventually the elongation ceases to be proportional
to the force. This is sometimes called the transition from
the elastic to the plastic regime.

phy 241 MenŽndez 25

 Chapter 2

In the problems we will consider, the spring constant K is a

quantity that must be obtained experimentally. On the other
hand, you might now conjecture that by carefully construct-
ing the potential energy function for our material we might
be able to derive the spring constant from the electrical forces
between atoms. Unfortunately, this is not an easy task,
because the interactions typically involve hundreds of
electrons and nuclei and because the problem must be
treated with quantum mechanics. It is only in the 1970Õs and
1980Õs that powerful methods have been developed which
allow the calculation of spring constants from first principles.
This is an active area of research, particularly at Arizona
State University.

The harmonic When HookeÕs law is inserted into NewtonÕs second law, we
obtain, as discussed in the previous chapter, the following
oscillator differential equation of motion:
d2 x K
2 =− x
dt m (1)

The solution to this equation is a function whose second

derivative is proportional to the function itself. We know
two types of elementary functions that have this property:
exponentials and the trigonometric functions sine and
cosine. As you will show in one of the homework problems,
exponentials cannot be a solution to Eq. (1). We can also
decide this issue on the basis of an experiment: if we
displace the mass, it will oscillate. The exponential function
is not oscillatory, but the trigonometric functions are. We
thus explore solutions of the form

x(t) = A cos ωt (2)

Notice that we have introduced two constants, the ampli-

tude A and the angular frequency ω. The amplitude A is
needed for two reasons: first, the function x(t) has units of
distance, while the trigonometric function cosine has no
units. Thus A has units of distance and insures the dimen-
sional consistency between the two sides of Eq. (2). On the
other hand, the function cosine runs between 1 and -1. Our

phy 241 MenŽndez 26

 Chapter 2

mass, however, will oscillate between any two values. As you

can easily show, these values will be precisely A and -A. The
angular frequency ω is introduced to compensate the units
of t in the argument of the cosine. This argument is in
radians. If we want time inside the argument, we must have
a factor with units of 1/time so that the units of the
argument cancel out. This factor is ω.

If you plug the proposed solution into Eq.(1), the following

relationship obtains:
K
ω2 =
m (3)

This means that Eq.(2) is indeed a solution to the spring

equation provided that we select a value of ω that satisfies
Eq. (3).

The meaning of ω

T
Displacement

Time
Figure 2 Oscillations of a mass attached to a spring

phy 241 MenŽndez 27

 Chapter 2

Fig. 2 displays the proposed solution, Eq. (2). As expected

for an oscillation, the function repeats itself after a certain
time. We call this time the period T . Its value can be
obtained from the condition

cos ω(t + T) = cos ωt (4)

which follows from the definition of period. This condition

has an infinite number of solutions, given by ωT = 2πn,
where n is any integer. But period is defined as the shortest
time for which the motion repeats itself, so that it corre-
sponds to n = 1. Thus the period is given by
2π
T =
ω (5)

When ω is large, the period is short, so that the oscillations

are very rapid. When ω is small, the period is long, so that
the oscillations are slow. Now consider Eq. (3). In view of
the meaning of ω, this equation means that the oscillations
will be faster the stronger the spring force, (large K) but
slower the larger the mass. This is a very important and
general result, which youÕll find again and again when we
discuss waves.

The angular frequency ω is always given by

something proportional to the strength of the
ÒrestoringÓ force divided by a quantity that
measures the inertia of the object.

The reason why ω is called angular frequency is that one can

define a quantity, known as the frequency ν, that measures
the number of oscillations per unit time. Thus the frequency
is the inverse of the period:
1 ω
ν = =
T 2π (6)

The angular frequency ω is therefore proportional to the

frequency ν. From Eq. (6) it is easy to see that the reason
for the word ÒangularÓ is that one oscillation per second
corresponds to ω = 2π 1 / s , whereby 2π is the number of

phy 241 MenŽndez 28

 Chapter 2

radians in a full circle. Of course, there are no physical angles

in the spring problem. Traditionally, however, the arguments
of the trigonometric functions are angles. This explains why
youÕll find the word ÒangleÓ quite frequently in the next few
sections, although - we insist - there are no geometrical
angles in this problem.

Completeness of the solution

It appears that we have found a solution to the equation of
motion for a mass attached to a spring. We now must
answer a fundamental question: is this solution complete? By
ÒcompleteÓ we mean the following: is Eq. (2) the solution to
the problem for any initial position and any initial velocity of
the mass? The form of Eq. (2) suggests that we might be in
trouble: the equation has two constants, A and ω. But ω is
Òused upÓ in Eq. (3), so that we are left with only one
ÒadjustableÓ constant with which we are supposed to
account for two numbers: the initial position and the initial
velocity.

LetÕs inspect the solution at t = 0. Eq. (2) gives x(0) = A cos

0 = A. This means that A is nothing but the initial position
of our mass. What is the velocity at t = 0? By differentiating
Eq. (2), we obtain v(t) = -ωA sin ωt. For t = 0, we obtain
v(0) = -ωA sin 0 = 0. Thus the initial velocity is zero! This
means that we have not yet found a complete solution to the
oscillating mass problem. Our solution only applies to cases
where the initial velocity is zero. For example, we might
stretch the spring to a certain position, wait for a moment
and release it. But if the initial velocity is different from
zero, Eq. (2) does not apply. What we have found is very
general: any solution to the equation of motion of a particle
must have two ÒadjustableÓ constants to account for the
arbitrary initial position and arbitrary initial velocity.

The complete solution to the spring problem

If you go back to our discussion leading to the proposal of
Eq. (2) as the solution of the problem, you may wonder why
we chose the cosine function instead of the sine function.
The fact is that both functions are solutions to our problem.

phy 241 MenŽndez 29

 Chapter 2

If you use x(t) = B sin ωt, youÕll find that this solution
satisfies Eq.(1) with the same condition specified in Eq. (3).
However, in this case we obtain x(0) = 0 and v(0) = ωB. So
the ÒcosineÓ solution gives a non-zero initial position and
zero initial velocity, whereas the ÒsineÓ solution gives zero
initial position and a non-zero initial velocity. How about
combining the two solutions? It turns out that the spring
equation is a linear differential equation with the following
property: if two arbitrarily different functions are solutions
to the equation, then their sum is also a solution. Hence we
propose as our complete solution

x(t) = A cos ωt + B sin ωt (7)

Linear homogenous differential equations

The spring equation belongs to a special kind of differential equations
that can be written as

dx d2 x dn x
A0 x + A 1 + A 2 2 + ... + A n n + ... = 0
dt dt dt
These equations have the following features:

¥ there are no terms independent of x, the function sought. Because

of this, the equations are called homogeneous. For example x - t3 = 0
dx
is not an a homogeneous equation, but x + 2 = 0 is
dt
homogeneous.

¥ There are terms no higher than the first power of x or its

derivatives. Because of this, the equation is said to be linear. For
2
dx dx ‘““
example, x + = 0 or x + ŠŒŒ = 0 are not linear, but
2
dt ‹ dt ’
the spring equation (1) is.

The fundamental property of linear differential equations is the superposi-

tion principle if we find two solutions, f1(t) and f2(t), any linear combina-
tion of these functions, B1 f1(t) + B2 f2(t), (where the BÕs are constants) is
also a solution of the equation. This is actually not a principle but an
easy theorem. You will verify the superposition theorem in one of the
end-of-chapter problems.

phy 241 MenŽndez 30

 Chapter 2

Eq. (7) has the two required ÒadjustableÓ constants to

account for the initial position and initial velocity. They are
given by

x(0) = A ; v(0) = ωB (8)

You will find in the literature several alternative ways of

writing Eq.(7). A popular one is to combine the two
trigonometric functions into one by using the identities
sin (α + β) = sinα cosβ + sinβ cosα
cos(α + β) = cosα cosβ − sinα sinβ (9)

Using this, we can rewrite Eq.(7) as

x(t) = A amp cos (ωt + α) (10)

Notice that this expression also has two ÒadjustableÓ

constants, A amp and α. Using the second identity in Eq. (9),
we find
A amp cos α = A = x 0
v0
− A amp sin α = B =
ω (11)

so that we can solve for the constants Aamp and α. We find

Š − v0
2
Š v0 ‘““ ‘““
A amp = x + ŒŒŒŒ ; α = arctan ŒŒŒŒ
2
0 ““ ““
‹ω ’ ‹ x0ω ’ (12)

It is a matter of convenience how we write our solution.

Eqs. (7) and (10) are mathematically identical. Eq. (10) has
the advantage that it contains a single trigonometric
function, which sometimes makes it easier to manipulate. In
addition, the coefficient Aamp has a simple physical interpre-
tation: it is the amplitude of the oscillation, i.e., the mass
oscillates between -A amp and A amp . On the other hand, the
ÒangleÓ α, called the Òphase angle,Ó does not represent a
geometrical angle. Its meaning is the following: at a time α/ω
before t=0, the mass would be at the maximum elongation
of the spring A amp , with zero instantaneous velocity. The

phy 241 MenŽndez 31

 Chapter 2

main disadvantage of Eq.(10) is that the determination of

the constants A amp and α from the initial conditions often
leads to mistakes. The problem lies in the ÒtanÓ function,
because there are two possible angles which have the same
tangent. You must find the correct one. You do this by going
back to Eqs. (11) and making sure that the signs you obtain
for the initial position and velocity match the signs of the
known initial conditions.

If you use Eq. (7), the expression is somewhat cumbersome,

but the constants A and B in are trivially determined from
the initial conditions. You should familiarize yourself with
the two alternative ways of expressing the solutions. Trigono-
metric functions play a fundamental role in the description
of wave phenomena. You should be able to manipulate them
without hesitation.

Same equation, In Eq. (1), the function we seek is the position as a function
of time of a mass attached to a spring. From the point of
same solution view of a mathematician, however, x(t) could represent
anything. Once we find a solution to the mass-spring
problem, this solution will be formally identical to the
solution of any other problem satisfying the same differential
equation. This problem need not have anything to do with
masses and springs: it could be the value of the Dow index,
the publicÕs opinion about Bill Clinton, or the number of
cockroaches in your garage. Anything whose second time
derivative is proportional to minus itself, will oscillate
exactly as indicated by Eq. (10). In physics, there are several
phenomena that lead to oscillations (thatÕs why we spend so
much time studying this phenomenon). A special case is that
of a pendulum, which weÕll discuss next.

An almost harmonic problem: the pendulum

Let us consider Fig. 3, where we show a mass m hanging
from a rod of length l. At a certain time t, we can decom-
pose the motion into a tangential and a normal component.
We arbitrarily select the bottom position as the origin from

phy 241 MenŽndez 32

 Chapter 2

which we measure angles, and we define as positive the

counter-clockwise direction. Of course, we could select the
zero and the positive direction in any way we want; no result
is affected by this choice. We define the arc displacement
function as s(t) = l θ(t). Because of our sign convention for
θ(t), s(t) is positive when the mass is to the right of the
vertical and negative when the mass is to the left of the
vertical. If we now write NewtonÕs second law for the
tangential motion, we obtain

− mg sin θ = m aT (13)

Notice the minus sign in front of the force term. The reason

000000000000000
111111111111111
for the minus sign is that the gravitational force acts to

111111111111111
000000000000000
reduce the arc when the mass is at the position of the figure.

111111111111111
000000000000000
0101
0101
01 θ l
0101
01 00
g 01 11
0101 00 11 m
0101 mg sin θ

01
Figure 3 Pendulum
d 2s
By definition, aT = dt 2
. Using the expression that gives the arc
as a function of the angle, and eliminating the mass (which
appears on the two sides of the equation), we obtain
d θ
2
ŒŠŒ g ‘““ sin θ
2 = −
dt ‹l ’ (14)

This equation is not equivalent to the spring equation,

phy 241 MenŽndez 33

 Chapter 2

because the second derivative of the angle is not proportion-

al to minus the angle but to minus the sine of the angle. In
fact, this is a very complicated equation, if you consider that
the sine of a quantity is given by an infinite series. However,
we can make an approximation for the case where the
angular displacement of the pendulum is small. (Not for the
drawing in Fig. 3!) With that restriction, we can approxi-
mate sin θ ∼ θ, in which case we obtain
d2 θ ŠŒŒ g ‘““
2 = − θ
dt ‹l ’ (15)

This equation is formally identical to Eq. (1) so that we can

immediately write the solution:

θ (t) = A amp cos ‡ ωt + α Ž (16)

with ω 2 = l . The constants Α αµπ and α are determined from

g

the initial conditions, which in this case are the initial angle
and the initial angular velocity. Notice that the quantity θ in
Eq. (16) corresponds to a real physical angle, namely the
angle between the pendulum and the vertical at any given
time. This means that the amplitude A amp has units of
angle. On the other hand, the ÒangleÓ α on the left-hand side
of the equation does not correspond to a physical angle: as
in the case of the mass and the spring, it is a parameter that
is determined from the initial conditions.

Damped harmonic The most important difference between ÒrealÓ harmonic

oscillators and the ÒtheoreticalÓ harmonic oscillators we
oscillators discussed in the previous section is the fact the oscillations
die out in real ones. A real harmonic oscillator, when left
alone, does not oscillate forever, as implied by our solutions
in Eqs. (7) or (10). The obvious reason for the this Òdamp-
ingÓ of the oscillations is the existence of friction. Friction
tends to stop the oscillation because friction forces always
oppose the motion. Thus the work they do is negative. This
work is equal to the change in energy of the mass attached
to the spring. This means that the energy of the mass

phy 241 MenŽndez 34

 Chapter 2

decreases and eventually becomes zero. Very often, the

damping is relatively slow: it takes many oscillations for the
mass to come to a final stop. In these cases, a given oscilla-
tion is very similar to the next, so that if you are interested
in these short times, you might neglect the effect of the
damping. Hence damping can often be ignored in many
problems involving the free oscillations of a mass-spring
system. However, damping becomes critically important
when you force the oscillations from outside the system, by
applying an external force. In these cases, the system reaches
an equilibrium between energy intake and energy loss only if
there is some friction. Friction is therefore essential to
discuss forced harmonic oscillators, the subject of the next
section. Mathematically, however, it is easier to start our
discussion with the free-oscillation case.

The origin of the friction force

Old-fashioned textbooks used to enumerate a long list of
different forces. The friction force was a prominent member
of this list. However, we know that there are only four
fundamental interactions in Nature, so that friction forces
must be a manifestation of one of those. The interactions
between human-size objects are electrical in nature: friction
must ultimately owe its origin to electrical interactions
between atoms. Unlike the case of HookeÕs law, however, it
is not possible to derive an expression for the friction force
based on general arguments. The reason is that many factors
will affect friction: the shape of the surface of the objects,
the atomic arrangement at the surface (which is often
different than in the bulk of the material), the temperature,
the strength of the atomic binding in a given material vis-a-
vis the strength of the interactions with other materials, etc,
etc. On the other hand, letÕs consider the form of the
solution to the equation of motion when we include friction
(you might wonder how can we possibly find a solution if we
donÕt know yet the form of the equation. The answer is that
we are in the Physics Department, not in the Math Depart-
ment. We can do experiments!). If you observe a real
harmonic oscillator, the oscillations will look as in Fig. 4.

phy 241 MenŽndez 35

 Chapter 2

Displacement

Time
Figure 4 The motion of a real, damped harmonic oscillator

Note that the amplitude of the oscillations diminishes as a

function of time. In fact, if you pass a function through the
maxima of the oscillations (this is the dashed line in the
figure) you will find that this function is usually proportional
to a simple exponential function of the form e -γt, where γ is
a constant. We might therefore adopt the point of view
Newton adopted when he discovered the gravitational law.
We might ask ourselves: what would be the simplest
expression for the friction force that leads to the solutions
depicted in Fig. 4? The answer is really very simple:

F friction = − λv, (17)

where λ is a constant and v is the velocity of the mass. This

equation has two appealing features: it automatically gives a

phy 241 MenŽndez 36

 Chapter 2

force that opposes the motion and, when included in the

equation of motion, preserves its linear character. However,
one should always be careful concerning the applicability of
this expression. In many common phenomena, such as the
friction experienced by an object travelling very fast in air,
the friction force is better approximated by a term propor-
tional to the square of the velocity. This, besides making it
even more difficult to design fast cars, leads to non-linear
equations of motion.

Solution to the damped oscillator problem

Let us now show how a solution like the one depicted in Fig.
4 is obtained when a friction force proportional to the
velocity is added to the spring [Link] equation of
motion for the mass is now a = (-Kx - λv)/m, which is
equivalent to
d2 x λ dx K
2 + + x=0
dt m dt m (18)

In view of Fig. 4, letÕs try a solution of the form

− γt
x(t) = A amp e cos (ωt + α) (19)

If we plug Eq. (19) into Eq. (18), we obtain after some

algebra and calculus the following (messy) equation:
− γt − γt
(γ 2 − ω 2 )A amp e cos(ωt + α) + 2γωA amp e sin (ωt + α)
λ − γt − γt
+ − γA amp e cos(ωt + α) − ωA amp e sin(ωt + α)
m
K − γt
+ A amp e cos(ωt + α) = 0
m (20)

This can only be satisfied if the terms that contain cosines

and sines vanish independently, leading to the following two
equations
γλ K
(γ 2 − ω 2 ) − + = 0
m m

λω
2γω − = 0
m (21)

phy 241 MenŽndez 37

 Chapter 2

We can now solve for ω and γ:

λ
γ =
2m
K
ω2 = −γ2
m (22)

The expression for γ makes sense: if the damping force

becomes zero (λ = 0), then γ = 0 and e-γt = 1, so that we
recover our solution without friction. The expression for ω
is somewhat surprising. It implies that damping affects not
only the amplitude of the motion but also its frequency. We
did not anticipate this effect when we proposed the solution.
The reason becomes quite obvious upon inspection of Eq.
(22). Even for small values of γ, we can see its effect on the
amplitude simply by waiting long enough. On the other
hand, if γ << ω, the frequency will be almost indistinguish-
able from the value without damping. Notice that our
solution requires ω 2 > 0. However, nothing prevents the
expression for ω 2 in Eq. (22) to become negative. If that
happens, our solution must be wrong. This is the case of
ÒoverdampingÓ: the friction is so strong that the mass
doesnÕt oscillate but approaches the equilibrium point
asymptotically. The solution is a combination of two
decaying exponentials, which is not surprising if you remem-
ber that exponentials and trigonometric functions are closely
related via complex numbers, which would appear in this
case because ω2 becomes negative.

Forced harmonic Let us now consider the most complicated case: a damped
harmonic oscillator subject to an external force. Let us
oscillations assume that the force is periodic and can be written as F(t)
= F0 cos ωft. Adding this term to Eq.(18) we obtain

d2 x λ dx K F0
2 + + x = cos ω f t
dt m dt m m (23)

This equation has a very important difference with Eq.(18).

The term of the right hand side is independent of x, so that
the equation is no longer homogeneous. Again we will find a
solution to this equation by observing experimental results.

phy 241 MenŽndez 38

 Chapter 2

They will convince you of a fundamental property of linear,

classical systems: if you excite them at a certain frequency,
the system responds at the same frequency. This is a result
worth remembering. It will play a fundamental role in our
analysis of wave motion in the coming chapters. If the
system responds at the same frequency, we can write the
solution to Eq.(23) as

x (t) = A cos ωf t + B sin ωf t (24)

If you plug this equation into Eq. (23) you obtain a lengthy
expression that contains cos ωft and sin ωft. The terms that
contain cos ω ft and the terms that contain sin ω ft must
satisfy the equation independently. This leads to two
equations whose solution is finally

F0 ω 20 − ω 2f
A=
m ω 2 − ω 2 2 + 4γ 2 ω 2
‡ 0 fŽ f

F0 2γω f
B=
m ω 2 − ω 2 2 + 4γ 2 ω 2
‡ 0 fŽ f (25)

where we have used ω 0 2 = Κ/ m and γ = λ/2m. In analogy

with the transformation leading to Eq. (10) from Eq. (7),
we can combine the two trigonometric functions in Eq. (24)
into a single expression of the form

x (t) = A amp cos ‡ ω f t + α Ž (26)

with

F0 / m
A amp = A2+ B2 = 2 2
‡ω 2
0 − ω 2f Ž + ‡ 2ω γ Ž
f

B 2ωγ
tan α = − =
A
‡ω 2
f − ω 20 Ž (27)

Several features of Eqs. (25) and (27) are worth discussing.

phy 241 MenŽndez 39

 Chapter 2

Notice that while the system oscillates at the external

frequency ω f, it somehow ÒremembersÓ that it ÒlikesÓ to
oscillate at the frequency ω0 of the free oscillations. So when
the external frequency ωf coincides with the natural frequen-
cy ω 0 , the amplitude A amp becomes very large. It actually
becomes infinity for vanishing damping and ω0 = ωf.

The other important feature about Eqs. (25) and (27) is

that the constants A and B (or A amp and α) have definite
expressions and are not dependent on the initial conditions.
(To avoid confusion, we should have used a different
notation for the constants that appear in free and forced
oscillations, since their physical meaning is very different.)
This means that our solution, Eq. (24) or Eq. (27), cannot
be the complete solution to our problem, since, as we
discussed before, there should be two ÒfreeÓ constants to
account for the initial position and velocity. Fortunately, the
linearity of the equation of motion provides an easy way out
of this problem. Eq. (23) is a linear inhomogeneous equation.
These equations satisfy a special version of the superposition
principle.

Superposition principle for linear inhomogeneous equations

Suppose that x1(t) is the solution to a linear inhomogeneous
equation where the external driving force is given by F1(t),
and x 2 (t) is the solution for the same equation when the
external driving force is given by F2(t). Then, if both driving
forces act simultaneously, so that the driving force becomes
F1(t) + F2(t), the sum x1(t) + x2(t) is the solution to the new
problem.

This theorem can be applied to the special case where one of

the external driving forces is zero. Of course, we have
discussed this case: free oscillations with damping, whose
solution is given by Eq. (19). Hence the complete solution
to the problem of a harmonic oscillator subject to an
external force F = F0 cos ωft is given by

phy 241 MenŽndez 40

 Chapter 2

x (t) = A amp cos ‡ ω f t + α Ž + B e − γt cos ‡ ωt + β Ž (28)

where Aamp and α are given by Eq. (27), ω and γ by Eq.(22),

and B and β are the ÒadjustableÓ constants that allow us to
accommodate any initial position and any initial velocity.

A very remarkable feature of Eq. (28) is what happens a long

time after the oscillation is started. Due to the exponential
in the second term, this term vanishes for very long times,
so that we are left with the first term, which has no Òinde-
pendentÓ constants. This means that two identical harmonic
oscillators subject to exactly the same external force but
starting with different initial conditions, will eventually
perform the same motion. In other words, the system
ÒforgetsÓ its initial conditions. This is clearly due to the
damping, for if γ = 0 then the second term never vanishes.
The result is easier to understand if you consider what
happens if a damped oscillator is left alone: it eventually
stops, no matter what the initial conditions are. This,
combined with the superposition principle, explains the
curious behavior of Eq. (28).

The first term of Eq.(28) is usually referred to as the

Òsteady stateÓ solution, whereas the second term is called
the Òtransient.Ó The meaning of this terminology is obvious
from our discussion above.

Resonance
When the exciting frequency ω f approaches the natural
frequency ω 0 , the system experiences a resonance. At
resonance, the amplitude of the oscillation, given by Eq.
(27) is dramatically enhanced. This means that the system
is able to absorb a large amount of energy from the external
driving force. The work done by this force is given by dW =
[F 0 cos (ω f t)] dx, where dx is the displacement of the
oscillator. If we divide both sides by dt we obtain the
instantaneous power given by
dW
P(t) = = F 0 cos ωf t v (t) (29)
dt

phy 241 MenŽndez 41

 Chapter 2

The velocity v(t) can be obtained by differentiating Eq. (24).

We obtain

P(t) = F 0 cos ωf t ω f B cos ω f t − ω f A sin ω f t

= F 0 ω f B cos 2 ω f t − F 0 ω f A cosωf t sin ω f t (30)

The function cos2 ωt is positive at all times, whereas the

product cos ω t sin ω t is positive and negative an equal
amount of time. During a period, the cos 2 term gives an
average value of 1 /2 , whereas the sin cos term averages to
zero. Hence the average power absorbed during a cycle is

γω 2f
2
F0
P= F0 ωf B =
1
2
m ‡ ω 2 − ω 2 Ž + 4γ 2 ω 2 (31)
0 f f

This function has a maximum for ωf = ω0, which means that

the maximum power absorption occurs for excitation at the
natural frequency of the system.

phy 241 MenŽndez 42

 Chapter 2

pendulum oscillates with an amplitude of 2¡.

Problems Express, as a function of time, a) its angular
displacement, b) its angular velocity, c) its angular
1. Show that x(t) = A e-at, where a is a real number, acceleration, d) its linear velocity, e) its centripetal
cannot be a solution to the harmonic oscillator acceleration and f) the tension on the string if the
equation. mass of the bob is 1 kg.

000
111 000
111
2. In problem 2, previous chapter, you studied
11. Find the period of oscillation for the mass in the
numerically the oscillations of a harmonic oscillator.

111
000 111
000
figure.
Find the analitycal solution to the same problem,
and plot the analytical solution on top of the
111
000 K1 K2
111
000
111
000 111
000
numerical solution.
m
3. A mass m = 2 kg is attached to a spring of
constant K = 8 N/m. The initial position is x(0) = 3
m and v(0) = 2 m/s.
111
000 111
000
a) By writing the solution as Eq. (7), determine all
12. Find the period of oscillation for the pendulum
constants in that expression.
in the figure.
b) By writing the solution as in Eq. (10), determine m
1
all constants in that expression.
c) Graph the two solutions in parts a) and b) to show
that they are identical.
l1 Pivot
4. Repeat Problem 3 with x(0) = -3 m.

5. Repeat Problem 3 with v(0) = - 2m/s

l2
6. Repeat Prob. 3 with x(0) = -3m and v(0)= -2 m/s.
m2
7. (Alonso 10.13) When a person of mass 60 kg
gets into a car, the center of gravity of the car lowers
0.3 cm. 13. Show that the current in the LC circuit of the
a) What is the elastic constant of the springs of the figure satisfies a harmonic oscillator equation with
2
car? ω2 = CL .
b) Given that the mass of the car is 500 kg, what is
its period of vibration whe it is empty and when the I
person is inside?

8. (Alonso 10.16) Find, for simple harmonic

motion, the values of ( x)ave and (x2)ave, where the
L
averages refer to time. Q1 Q2
9. (Alonso 10.17) Find the average values of the
C C
kinetic and potential energies in simple harmonic -Q 1 -Q 2
motion relative to a) time and b) position.

10. (Alonso 10.20) A simple pendulum whose

length is 2 m is in a place where g = 9.80 m/s 2. The

phy 241 MenŽndez 43

 Chapter 2

14. Using Eq. (19), find an expression for the initial

position and velocities in the case of a dampled
harmonic oscillator. If the oscillator of Problem 3
has a damping constant λ = 0.2 kg/s, find a
numerical expresion for the position of the mass as a
function of time. Graph this function.

15. Prove the theorem of superposition for linear

inhomogeneous equations. Show with an example
that the theorem is not valid for non-linear
equations.

16. Complete the derivation leading to Eq. (25).

17. Derive Eq. (31) by computing the averages in a

rgurous way and plot the average power as a
function of the frequency ωf.

18. For the oscillator in the previous problem, find

the average power given to the oscillator by the
friction force. Show that this power equals minus
the power given by Eq. (31) . Discuss.

19. Compute and plot the instantaneous total

energy of the oscillator driven by an external force.
Does your result contradict your conclusions in
Problem 18?

20. (Alonso 10.31) Find the limiting values of the

amplitude and the phase of a forced damped
oscillator when a) ωf is much smaller than ω0 and b)
when ω f is much larger than ω 0 . Determine the
dominant factors in each case.

phy 241 MenŽndez 44